Re: Crash in pgCrypto?

David Fetter <david@fetter.org> writes:
> It's not quite that simple.  Let's say you're *developing* a module.
> I don't see any way to play with it in the separate module proposal,
> where I *do* see a whole extra non-orthogonal feature where none is
> needed.

The claim that no new feature is needed is complete rubbish.  The
*main* thing that we need to get out of a module concept is to have
pg_dump know that it should not dump objects that are part of a
module (at least in the default case).  That can't be the behavior
for schemas.

You could imagine implementing modules as specially marked schemas,
perhaps, but I don't see any particular advantage to that.  In
particular, I don't want to force people to play around with
search_path in order to use modules.

> Here's how what I'm proposing would work:

> 1.  Create a way for schemas themselves to depend on other schemas,
> *not* on the stuff inside.

That does not actually solve any problem we need solved.
        regards, tom lane